Table 1 Summary of discussion themes and recommendations made by the panel

1. Standardization

●In order to compare CAIM with conventional therapies, the same basic economic evaluation framework and methods must be used.

2. The question, audience, and perspective

●Given the high level of self-referral and out-of-pocket payment the patients’ perspective as to important health outcomes and relevant costs should be incorporated into CAIM economic evaluations either as a separate analysis or as part of the societal perspective when that perspective is used.

3. Analytic methods

●Both the treatment and usual care arms of CAIM trials should be described in detail given the heterogeneity of treatment across practitioners.

●To capture the preventive nature of CAIM, longer follow ups and/or modeling should be used to capture impacts beyond the treatment phase.

4. Outcomes

●Given the tendency for CAIM to treat the whole person, rather than only the targeted condition, measure and report a wide range of outcomes, and/or to use a summary measure of overall health that could capture the full range such as quality-adjusted life-years (QALYs).

●If multiple outcomes are used, consider reporting these as a cost-consequence analysis.

5. Costs

●Given that CAIM is not often covered by health insurance (whose claims databases are the source of much of the cost data available to studies of conventional medicine) and since the relative quality of other sources of cost data (e.g., practitioner records or patient self-report) is not always clear, both sets of costs could be reported.

●Since CAIM therapies tend to treat the whole person they may have cost impacts beyond the targeted disease. Therefore, it is recommended that overall healthcare costs be reported, either in addition to or instead of disease-specific costs.

6. Implementation

●Highlighting studies where (and the circumstances under which) CAIM can both improve health outcomes and reduce costs (rather than increase costs, even if still considered cost effective) might gain it more attention from policy makers.

7. Generalizability

●The first recommendation made under theme 3 above is reiterated here; this information is also needed to enable study results to be adapted to other settings.

●As compared to conventional therapies, CAIM therapies might require more sensitivity analyses to test the robustness of results given the large variability in treatment methods, dosages, and pricing.

●Given that patient beliefs can have a large impact on outcomes, especially in CAIM, these beliefs should be measured and reported to allow comparisons to other populations.